RatingLift helps local businesses (restaurants, salons, dental and medical clinics, auto repair, HVAC and plumbing, gyms, retail, hotels, cafes) earn more genuine reviews and plan the rating math: compute how many new reviews lift the average to a target and how long that takes at the current pace, plan monthly review requests by channel (text, email, QR card, in-person, receipt link) with built-in typical response rates and a projected rating path, draft compliant request messages with an opt-out line and a compliance checklist, compute send timing and a weekly schedule, compare review tools by cost per new review, and score the business profile with fixes ranked by impact. It enforces platform rules: no incentives, no review gating, ask every customer. Planning and drafting only - it never sends messages, never posts or fetches reviews, and gives no legal or financial advice.
MCP endpoint (Streamable HTTP): POST https://rating-lift.magicteams.ai/mcp
Find RatingLift in the ChatGPT Apps, Claude Connectors, and Meta Muse directories — or connect the MCP URL directly.
| Tool | What it does |
|---|---|
rating_math | Compute how many new reviews (at a given average, default 5 stars) a business needs to lift its average rating to a target, with a table of common targets, the count for the displayed one-decimal rating, and months needed at the current review velocity. Use when someone asks how many reviews it takes to reach a rating. Do NOT use to plan which channels to ask through; use plan_review_requests. |
plan_review_requests | Plan monthly review requests for a business type: built-in typical ask and response rates per channel (sms, email, qr_card, in_person, receipt_link) give expected new reviews and messaging cost per channel, a recommended mix, and a projected rating path over the coming months when you give the current rating and count. Use when deciding how to ask. Do NOT use to write the message text; use draft_review_request. |
draft_review_request | Draft a compliant review-request message for sms, email, a printed card or an in-person script: business name, link placeholder, service, timing after the visit and tone, with an opt-out line and a compliance checklist (no incentive, no gating, honest review, ask everyone). Rejects supplied text that offers an incentive or asks only happy customers. Draft only; nothing is sent. Do NOT use for when to send; use build_request_timing. |
build_request_timing | Compute when to send review requests for a service type: send delay after the visit and best time of day (shifted out of quiet hours), one reminder after N days, opt-out line, and a weekly send schedule by service day with closed days skipped. Schedule only; nothing is scheduled or sent. Do NOT use for message wording; use draft_review_request. |
compare_review_tools | Compare review-request tools you are considering (name, monthly price, locations covered, included messages, per-message cost) against your expected new reviews and message volume: monthly cost with overage, cost per new review, cost per location, period total, ranked cheapest per review first. Omit options for a reference table of typical US tiers (typical estimate). Use when choosing software. Do NOT use to forecast review volume; use plan_review_requests. |
audit_review_profile | Score a business profile from facts you provide (rating, review count, days since last review, owner response rate, photo count, categories, hours, duplicate listings, description, Q&A) and return a 0-100 score with fixes ranked by impact and the reason for each. Uses only what you type; fetches nothing. Use for a listing health check. Do NOT use for how many reviews a rating target needs; use rating_math. |
Free. Computed per request; nothing stored.
llms.txt · agent.json · UCP profile · sitemap